Docker Model Runner

Run AI Models Locally. Full Control and Zero Hassle

Local-first LLM inference made easy

Docker Model Runner makes it easy to test and run AI models locally using familiar Docker CLI commands and tools. It works with any OCI-compliant registry, including Docker Hub, and supports OpenAI’s API for quick app integration.

Cut down on token costs, keep your data private, and stay in full control!
docker model run

Run popular AI models locally with Docker simplicity

Models ready to go

Explore a curated catalog of open-source AI models on Docker Hub or pull directly from Hugging Face.

Full Local control

Run LLMs locally and take full ownership of your AI workflows and data privacy.

Native GPU acceleration

Maximize performance on your local machine with fast, efficient inference.

Easy app integration

Send inference requests via OpenAI’s API, perfect for fast, scalable app integration.

Seamless Docker integration

Use with popular Docker tools, including Docker Compose, Testcontainers, and Offload for easy scaling from local to CI/CD.

Portable by design, easy to share

Package, manage, and share models to any OCI-compliant registry such as Docker Hub.

Built for teams and enterprise

Securely manage access and control for your models as OCI-artifacts in Docker Hub.

Use it where you build AI apps

Connect with frameworks such as Spring AI, LangChain, OpenWebUI, and more using the Docker tooling you already know and trust.

Hassle-free local inference starts here

Run LLMs locally with Docker Model Runner. Cut token cost, keep control, and use the tools you already know.